mpn_rshift / mpn_lshift bug on m68000

[Kevin Ryde]

Patrick Pelissier <Patrick.Pelissier at loria.fr> writes:
>
> PS: Another bug is that you define in gmp-impl.h:

[Separate messages for separate matters please.]

>   but mc68000 doesn't have native Floating Point Unit, so floats could
> be implement by software library, which could be not in IEEE
> format. Maybe it would be better to detect the format in configure?

Yep, this will be done for the next release.  The tests are in fact
already in the 4.1.2 configure, but not used.

PalmOS is one possible case of this.  I believe it defines its own
format software float, but last time I looked their gcc port wasn't
ready to make direct use of it.  Perhaps that's changed by now.